The ACD Method

[Maverick74]

Best FX trade going right now.

 

[DT3]

There are two roadblocks to prediction that make things really really messy. The first is variance ("a funny thing happened on the way to the forum") and the second is time. These two variables have a way of spoiling the best of parties. Because of these variables, they prohibit prediction. Because without these variables in the model, you are nothing more then the random drunk leaving the bar stumbling with no particular place to go. We have to accept this. There is no way around it. So what ACD does, or at least in my ACD playland, I incorporated them into my model. If you can't beat them, join them right? So I did something funny on my way to the forum, I said the hell with price, let me predict variance and time and I'll invite price along for desert later. The result was a much greater understanding of the market then staring at price moving up and down the way a cat's eyes follow the ball on the end of a string shocked as it moves around in space.

Once you understand the limitations of your model, you can move forward in peace. All is well with the world. The biggest part in building a model or a methodology is understanding everything that it is and everything that it is not. The model if built correctly should explain the market as it is, not as you want it to be. It should help bring clarity, instill discipline, minimize emotion and provide an objective path forward that ultimately will determine value.

Mav this post went over my head. I think I understand what you are alluding to with time, time stops,time confirmations. But I'm not getting the variance part.

Example: You decide to buy X at 100. You think it will go to 110 and put in a stop at 98. What if X goes to 95 first and then goes to 110. Was your "prediction" right? It went to 110 right? Well no, it went to 95 first, you lose 2 pts...you lose...variance wins.

Round 2: You say the hell with stops then, stops are for suckers (or so you read on ET). So you buy X at 100 with a target of 110. No stop this time. X goes down to 90....rebounds back to 100....then continues on to 110....hits your target. Did you win? No. All you did was bet on variance. You bought X at 100 and it went 10 pts below and 10 pts above....in other words, completely random. You don't win anything. Although I'm sure 100 posts of bragging will follow. LOL.

Round 3: You buy X at 100...no stop. This time it goes no where for 6 months but finally hits 110. You sell. Did you win? Again...no. Let's say over that same time period the index was up 20%, X was up 10% and the index exhibited half the volatility as X. The index in this case represents your opportunity cost which in economics, we count as a real expense.

Are we using ACD in the above examples? If we are using ACD as a guide wouldn't the fact that price went underneath our stop be the sign that we are wrong. In other words we shouldn't be placing stops at arbitrary levels, they should be at a level where if it hits we know we're wrong and need to get out.

 

[Robert Yanks]

Is Emerging Markets (EEM) getting ready to form a higher high?

My 30 day confirmed on 4/22/15 and has stayed strong the whole pullback. The 5 day is at 7 and would ordinarily work off a few now.

There was a Monthly A Down on 6/5 but you can see the pattern held in place, and the Quarterly A Down at 39.15 has held nicely.

The higher highs and higher lows are ……. Obvious

 

[wow12]

DT3 when Mav referring to variance I think he’s referring to risk.


Since Mav been posting again (thanks again Mav) few things have become more apparent especially my understanding about the markets and since those posting few ideas have developed which I’ll refer to as hypothesis 1, 2 and 3 (ironically all this has been staring in front of my face)


Hypothesis 1: the market always seeks risk, therefore designed to screw most amount of traders efficiently as possible.( pretty efficient in that respect)


Hypothesis 2: Markets are not efficient in finding value refer to hypothesis 1


Hypothesis 3: when one argues that you can eliminate risk through hedging not entirely true but I do believe you can control it.


Ok guys these are some ideas that I have come up with regards to the market and my understanding, let me know what you guys think (Mav deffo want to hear your views see if i'm on the right track)

 

[Robert Yanks]

Variance is a measure of distribution. It's usually synonymous with risk and denoted as "sigma^2" or it's little brother sigma "standard deviation". One way to think about it is, it's working counter to what you are trying to do. You are trying to buy X at 100 and sell it at 110 and you prefer the optimal path to get there which is a smooth slow moving straight line. Sigma is there to make sure it doesn't happen. It's going to make the ride bumpy and violent in hopes of getting you off the bus before it arrives at its destination.

Example: You decide to buy X at 100. You think it will go to 110 and put in a stop at 98. What if X goes to 95 first and then goes to 110. Was your "prediction" right? It went to 110 right? Well no, it went to 95 first, you lose 2 pts...you lose...variance wins.

Round 2: You say the hell with stops then, stops are for suckers (or so you read on ET). So you buy X at 100 with a target of 110. No stop this time. X goes down to 90....rebounds back to 100....then continues on to 110....hits your target. Did you win? No. All you did was bet on variance. You bought X at 100 and it went 10 pts below and 10 pts above....in other words, completely random. You don't win anything. Although I'm sure 100 posts of bragging will follow. LOL.

Round 3: You buy X at 100...no stop. This time it goes no where for 6 months but finally hits 110. You sell. Did you win? Again...no. Let's say over that same time period the index was up 20%, X was up 10% and the index exhibited half the volatility as X. The index in this case represents your opportunity cost which in economics, we count as a real expense.

So you can see there is a lot of ways to randomly capture this trade and scream out winner winner chicken dinner! The key to this trade is to be able to capture that 10 pt move...with controlled risk that does not get triggered in a meaningful time frame and at the same time exceeding your opportunity cost. And there in lies the rub. If you irresponsibly ignore variance and time, you might give someone the impression you are a better trader then you really are. But include those variables and suddenly gravity takes hold. Nobody said pimpin was easy.

I just noticed that "Mav74's Underground ACD Ebook" has a new chapter: Variance. Cool

 

[eurusdzn]

I would like to ask what factors people put int here models.
I have 60 day bars in mine, along with weekly and daily. So easy to see a mean reverting
asset. Failed and broken extreme levels seem to be more significant and meaningful with 60
day bars.
Volatility and noise modeling , well, they require some work. Seems that pure ACD would require defining noise as a custom fraction of the Aup and Adown levels. What I really dont know Mav is if you mean that volatility that occurs within a defined noise area is ignored and to wait until price commits to a direction beyond the noise. Patience. Am i off base here?
Thanks.

 

[ignl]

Here is the quote from the book "Mapping the mind" which made me stop a bit and think. Intuitively this is probably well known for traders but it is still very interesting from neuroscience perspective. This is my contribution to share something interesting

Hence we arrive in the world with a set of hypotheses which we update only if the incoming evidence conflicts with them. We also have a build-in 'confirmation bias' that encourages us to be alert to the evidence that supports our prejudice than to that which contradicts it. So, for example, if we come across unpalatable truths about someone we want to admire, we override the impact of it by sheer emotional force. Psychologists at Emory University demonstrated this in a sample of committed Democrats and Republicans during three months prior to he U.S. Presidential election of 2004. The Democrats and Republicans were each given information about their chosen candidate which - had they been operating in a pure reasonable way - would have given them second thoughts about their preference. While they took in the information the researchers scanned their brains with fMRI to see what was happening in them. What they found was that the frontal lobe areas which are known to be involved in reasoning were strangely quiet. Instead increased activation was seen in a network of emotion circuits, including those activated sadness, disgust, and - most tellingly, perhaps - conflict. The subjects' brains looked for all the world as though they were fighting the incoming information, struggling to block it out or minimize its impact. The subjects were then invited to comment on what they had learned and more or less all of them reported that their essential preference was unchanged, and that the information they had received was less important than it seemed or - by various acts of mental athletics - actually supportive of their views. Having arrived at their convoluted conclusions, the activity in subjects' brains altered. Now the areas that lit up were those in the reward circuit. Not only, it seems, do we find ways to support our prejudices but, having done so, we reward ourselves for it!
There is an important difference between these higher illusion-creating mechanisms and those that make us see sticky-out noses on concave faces. It is that they are much more malleable. Try as you might, you will not get rid of the illusory grey square in Hermann's grid. But you can work at the illusion of guilty man, or illusion of the 'nice, even, winning type' string of lottery numbers. You can chip away at your political prejudices; force yourself to asses rationally evidence you would normally fight to dismiss.
Thinking alters thinking. We can actually change the structure and activity in our brains, just by deciding to. It is the greatest accomplishment of our species.

 

[wow12]

I would like to ask what factors people put int here models.
I have 60 day bars in mine, along with weekly and daily. So easy to see a mean reverting
asset. Failed and broken extreme levels seem to be more significant and meaningful with 60
day bars.
Volatility and noise modeling , well, they require some work. Seems that pure ACD would require defining noise as a custom fraction of the Aup and Adown levels. What I really dont know Mav is if you mean that volatility that occurs within a defined noise area is ignored and to wait until price commits to a direction beyond the noise. Patience. Am i off base here?
Thanks.

I think variance helps measure volatility but I'm not 100% certain on that (better to wait for Mav).

 

[samuel11]

I think variance helps measure volatility but I'm not 100% certain on that (better to wait for Mav).

Volatility is the square root of variance.

 

[kinggyppo]

Mav this post went over my head. I think I understand what you are alluding to with time, time stops,time confirmations. But I'm not getting the variance part.



Are we using ACD in the above examples? If we are using ACD as a guide wouldn't the fact that price went underneath our stop be the sign that we are wrong. In other words we shouldn't be placing stops at arbitrary levels, they should be at a level where if it hits we know we're wrong and need to get out.

regarding the first part not exactly sure how to make this less confusing, there some advanced concepts such as variance thrown in. Variance is different from volatility which is another can of worms. I think what he is saying is that variance is the idea that stocks move around. Looking at a one year chart of MSFT it has a high of 50.05 and a low
40.12, so not a ton of movement. The historical volatility of MSFT which also bounces around based on what's going on with the stock is around 20. This means that on average MSFT will move up or down 20% annually. hopefully I haven't confused you more. In general the more a stock moves around the higher the vol.